Confused about SnapMapper::get_prefix

Confused about SnapMapper::get_prefix

[Ding Dinghua]

Hi all:
      I don't understand why SnapMapper::get_prefix   static_cast snap
to unsigned:

string SnapMapper::get_prefix(snapid_t snap)
{
  char buf[100];
  int len = snprintf(
    buf, sizeof(buf),
    "%.*X_", (int)(sizeof(snap)*2),
    static_cast<unsigned>(snap));
  return MAPPING_PREFIX + string(buf, len);
}

Will this limit snapshot count in pool to 2^32 -1 ?

Could anyone clarify ? Thanks

-- 
Ding Dinghua

Re: Confused about SnapMapper::get_prefix

[Gregory Farnum]

On Tue, Feb 3, 2015 at 4:12 AM, Ding Dinghua <dingdinghua85@gmail.com> wrote:
> Hi all:
>       I don't understand why SnapMapper::get_prefix   static_cast snap
> to unsigned:
>
> string SnapMapper::get_prefix(snapid_t snap)
> {
>   char buf[100];
>   int len = snprintf(
>     buf, sizeof(buf),
>     "%.*X_", (int)(sizeof(snap)*2),
>     static_cast<unsigned>(snap));
>   return MAPPING_PREFIX + string(buf, len);
> }
>
> Will this limit snapshot count in pool to 2^32 -1 ?
>
> Could anyone clarify ? Thanks

I think the code base is a little confused about whether snaps should
be 32 or 64 bits in various places. :( That said, the size of unsigned
can vary across architectures, so this should probably be sized more
explicitly as whatever it's supposed to be on the disk...

Re: Confused about SnapMapper::get_prefix

[Samuel Just]

It looks like snapid_t is a uint64_t, but snprintf expects an unsigned there.
-Sam

On Tue, Feb 3, 2015 at 9:15 AM, Gregory Farnum <greg@gregs42.com> wrote:
> On Tue, Feb 3, 2015 at 4:12 AM, Ding Dinghua <dingdinghua85@gmail.com> wrote:
>> Hi all:
>>       I don't understand why SnapMapper::get_prefix   static_cast snap
>> to unsigned:
>>
>> string SnapMapper::get_prefix(snapid_t snap)
>> {
>>   char buf[100];
>>   int len = snprintf(
>>     buf, sizeof(buf),
>>     "%.*X_", (int)(sizeof(snap)*2),
>>     static_cast<unsigned>(snap));
>>   return MAPPING_PREFIX + string(buf, len);
>> }
>>
>> Will this limit snapshot count in pool to 2^32 -1 ?
>>
>> Could anyone clarify ? Thanks
>
> I think the code base is a little confused about whether snaps should
> be 32 or 64 bits in various places. :( That said, the size of unsigned
> can vary across architectures, so this should probably be sized more
> explicitly as whatever it's supposed to be on the disk...

Re: Confused about SnapMapper::get_prefix

[Samuel Just]

Should probably be cast to long unsigned with lX conversion specifier?
-Sam

On Tue, Feb 3, 2015 at 9:21 AM, Samuel Just <sam.just@inktank.com> wrote:
> It looks like snapid_t is a uint64_t, but snprintf expects an unsigned there.
> -Sam
>
> On Tue, Feb 3, 2015 at 9:15 AM, Gregory Farnum <greg@gregs42.com> wrote:
>> On Tue, Feb 3, 2015 at 4:12 AM, Ding Dinghua <dingdinghua85@gmail.com> wrote:
>>> Hi all:
>>>       I don't understand why SnapMapper::get_prefix   static_cast snap
>>> to unsigned:
>>>
>>> string SnapMapper::get_prefix(snapid_t snap)
>>> {
>>>   char buf[100];
>>>   int len = snprintf(
>>>     buf, sizeof(buf),
>>>     "%.*X_", (int)(sizeof(snap)*2),
>>>     static_cast<unsigned>(snap));
>>>   return MAPPING_PREFIX + string(buf, len);
>>> }
>>>
>>> Will this limit snapshot count in pool to 2^32 -1 ?
>>>
>>> Could anyone clarify ? Thanks
>>
>> I think the code base is a little confused about whether snaps should
>> be 32 or 64 bits in various places. :( That said, the size of unsigned
>> can vary across architectures, so this should probably be sized more
>> explicitly as whatever it's supposed to be on the disk...

Re: Confused about SnapMapper::get_prefix

[Ding Dinghua]

I think so, snap_id is defined and used as uint64_t in ceph, but here
static_cast may introduce bug, since two snap_id may get the same
prefix, then same key in snap_mapper.

diff --git a/src/osd/SnapMapper.cc b/src/osd/SnapMapper.cc
index 315e2e2..27cc2b7 100644
--- a/src/osd/SnapMapper.cc
+++ b/src/osd/SnapMapper.cc
@@ -72,8 +72,7 @@ string SnapMapper::get_prefix(snapid_
t snap)
   char buf[100];
   int len = snprintf(
     buf, sizeof(buf),
-    "%.*X_", (int)(sizeof(snap)*2),
-    static_cast<unsigned>(snap));
+    "%.*llX_", (int)(sizeof(snap)*2), snap);

2015-02-04 1:25 GMT+08:00 Samuel Just <sam.just@inktank.com>:
> Should probably be cast to long unsigned with lX conversion specifier?
> -Sam
>
> On Tue, Feb 3, 2015 at 9:21 AM, Samuel Just <sam.just@inktank.com> wrote:
>> It looks like snapid_t is a uint64_t, but snprintf expects an unsigned there.
>> -Sam
>>
>> On Tue, Feb 3, 2015 at 9:15 AM, Gregory Farnum <greg@gregs42.com> wrote:
>>> On Tue, Feb 3, 2015 at 4:12 AM, Ding Dinghua <dingdinghua85@gmail.com> wrote:
>>>> Hi all:
>>>>       I don't understand why SnapMapper::get_prefix   static_cast snap
>>>> to unsigned:
>>>>
>>>> string SnapMapper::get_prefix(snapid_t snap)
>>>> {
>>>>   char buf[100];
>>>>   int len = snprintf(
>>>>     buf, sizeof(buf),
>>>>     "%.*X_", (int)(sizeof(snap)*2),
>>>>     static_cast<unsigned>(snap));
>>>>   return MAPPING_PREFIX + string(buf, len);
>>>> }
>>>>
>>>> Will this limit snapshot count in pool to 2^32 -1 ?
>>>>
>>>> Could anyone clarify ? Thanks
>>>
>>> I think the code base is a little confused about whether snaps should
>>> be 32 or 64 bits in various places. :( That said, the size of unsigned
>>> can vary across architectures, so this should probably be sized more
>>> explicitly as whatever it's supposed to be on the disk...



-- 
Ding Dinghua

Re: Confused about SnapMapper::get_prefix

[Ding Dinghua]

Ping..

I think so, snap_id is defined and used as uint64_t in ceph, but here
static_cast may introduce bug, since two snap_id may get the same
prefix, then same key in snap_mapper.

diff --git a/src/osd/SnapMapper.cc b/src/osd/SnapMapper.cc
index 315e2e2..27cc2b7 100644
--- a/src/osd/SnapMapper.cc
+++ b/src/osd/SnapMapper.cc
@@ -72,8 +72,7 @@ string SnapMapper::get_prefix(snapid_
t snap)
   char buf[100];
   int len = snprintf(
     buf, sizeof(buf),
-    "%.*X_", (int)(sizeof(snap)*2),
-    static_cast<unsigned>(snap));
+    "%.*llX_", (int)(sizeof(snap)*2), snap);

2015-02-04 12:02 GMT+08:00 Ding Dinghua <dingdinghua85@gmail.com>:
> I think so, snap_id is defined and used as uint64_t in ceph, but here
> static_cast may introduce bug, since two snap_id may get the same
> prefix, then same key in snap_mapper.
>
> diff --git a/src/osd/SnapMapper.cc b/src/osd/SnapMapper.cc
> index 315e2e2..27cc2b7 100644
> --- a/src/osd/SnapMapper.cc
> +++ b/src/osd/SnapMapper.cc
> @@ -72,8 +72,7 @@ string SnapMapper::get_prefix(snapid_
> t snap)
>    char buf[100];
>    int len = snprintf(
>      buf, sizeof(buf),
> -    "%.*X_", (int)(sizeof(snap)*2),
> -    static_cast<unsigned>(snap));
> +    "%.*llX_", (int)(sizeof(snap)*2), snap);
>
> 2015-02-04 1:25 GMT+08:00 Samuel Just <sam.just@inktank.com>:
>> Should probably be cast to long unsigned with lX conversion specifier?
>> -Sam
>>
>> On Tue, Feb 3, 2015 at 9:21 AM, Samuel Just <sam.just@inktank.com> wrote:
>>> It looks like snapid_t is a uint64_t, but snprintf expects an unsigned there.
>>> -Sam
>>>
>>> On Tue, Feb 3, 2015 at 9:15 AM, Gregory Farnum <greg@gregs42.com> wrote:
>>>> On Tue, Feb 3, 2015 at 4:12 AM, Ding Dinghua <dingdinghua85@gmail.com> wrote:
>>>>> Hi all:
>>>>>       I don't understand why SnapMapper::get_prefix   static_cast snap
>>>>> to unsigned:
>>>>>
>>>>> string SnapMapper::get_prefix(snapid_t snap)
>>>>> {
>>>>>   char buf[100];
>>>>>   int len = snprintf(
>>>>>     buf, sizeof(buf),
>>>>>     "%.*X_", (int)(sizeof(snap)*2),
>>>>>     static_cast<unsigned>(snap));
>>>>>   return MAPPING_PREFIX + string(buf, len);
>>>>> }
>>>>>
>>>>> Will this limit snapshot count in pool to 2^32 -1 ?
>>>>>
>>>>> Could anyone clarify ? Thanks
>>>>
>>>> I think the code base is a little confused about whether snaps should
>>>> be 32 or 64 bits in various places. :( That said, the size of unsigned
>>>> can vary across architectures, so this should probably be sized more
>>>> explicitly as whatever it's supposed to be on the disk...
>
>
>
> --
> Ding Dinghua



-- 
Ding Dinghua

Re: Confused about SnapMapper::get_prefix

[Sage Weil]

This looks right to me.  Sam?

sage


On Sun, 8 Feb 2015, Ding Dinghua wrote:

> Ping..
> 
> I think so, snap_id is defined and used as uint64_t in ceph, but here
> static_cast may introduce bug, since two snap_id may get the same
> prefix, then same key in snap_mapper.
> 
> diff --git a/src/osd/SnapMapper.cc b/src/osd/SnapMapper.cc
> index 315e2e2..27cc2b7 100644
> --- a/src/osd/SnapMapper.cc
> +++ b/src/osd/SnapMapper.cc
> @@ -72,8 +72,7 @@ string SnapMapper::get_prefix(snapid_
> t snap)
>    char buf[100];
>    int len = snprintf(
>      buf, sizeof(buf),
> -    "%.*X_", (int)(sizeof(snap)*2),
> -    static_cast<unsigned>(snap));
> +    "%.*llX_", (int)(sizeof(snap)*2), snap);
> 
> 2015-02-04 12:02 GMT+08:00 Ding Dinghua <dingdinghua85@gmail.com>:
> > I think so, snap_id is defined and used as uint64_t in ceph, but here
> > static_cast may introduce bug, since two snap_id may get the same
> > prefix, then same key in snap_mapper.
> >
> > diff --git a/src/osd/SnapMapper.cc b/src/osd/SnapMapper.cc
> > index 315e2e2..27cc2b7 100644
> > --- a/src/osd/SnapMapper.cc
> > +++ b/src/osd/SnapMapper.cc
> > @@ -72,8 +72,7 @@ string SnapMapper::get_prefix(snapid_
> > t snap)
> >    char buf[100];
> >    int len = snprintf(
> >      buf, sizeof(buf),
> > -    "%.*X_", (int)(sizeof(snap)*2),
> > -    static_cast<unsigned>(snap));
> > +    "%.*llX_", (int)(sizeof(snap)*2), snap);
> >
> > 2015-02-04 1:25 GMT+08:00 Samuel Just <sam.just@inktank.com>:
> >> Should probably be cast to long unsigned with lX conversion specifier?
> >> -Sam
> >>
> >> On Tue, Feb 3, 2015 at 9:21 AM, Samuel Just <sam.just@inktank.com> wrote:
> >>> It looks like snapid_t is a uint64_t, but snprintf expects an unsigned there.
> >>> -Sam
> >>>
> >>> On Tue, Feb 3, 2015 at 9:15 AM, Gregory Farnum <greg@gregs42.com> wrote:
> >>>> On Tue, Feb 3, 2015 at 4:12 AM, Ding Dinghua <dingdinghua85@gmail.com> wrote:
> >>>>> Hi all:
> >>>>>       I don't understand why SnapMapper::get_prefix   static_cast snap
> >>>>> to unsigned:
> >>>>>
> >>>>> string SnapMapper::get_prefix(snapid_t snap)
> >>>>> {
> >>>>>   char buf[100];
> >>>>>   int len = snprintf(
> >>>>>     buf, sizeof(buf),
> >>>>>     "%.*X_", (int)(sizeof(snap)*2),
> >>>>>     static_cast<unsigned>(snap));
> >>>>>   return MAPPING_PREFIX + string(buf, len);
> >>>>> }
> >>>>>
> >>>>> Will this limit snapshot count in pool to 2^32 -1 ?
> >>>>>
> >>>>> Could anyone clarify ? Thanks
> >>>>
> >>>> I think the code base is a little confused about whether snaps should
> >>>> be 32 or 64 bits in various places. :( That said, the size of unsigned
> >>>> can vary across architectures, so this should probably be sized more
> >>>> explicitly as whatever it's supposed to be on the disk...
> >
> >
> >
> > --
> > Ding Dinghua
> 
> 
> 
> -- 
> Ding Dinghua
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> 

Re: Confused about SnapMapper::get_prefix

[Samuel Just]

I think that's right.  Do you want to create a pull request against hammer?
-Sam

On Sun, Feb 8, 2015 at 7:07 AM, Sage Weil <sweil@redhat.com> wrote:
> This looks right to me.  Sam?
>
> sage
>
>
> On Sun, 8 Feb 2015, Ding Dinghua wrote:
>
>> Ping..
>>
>> I think so, snap_id is defined and used as uint64_t in ceph, but here
>> static_cast may introduce bug, since two snap_id may get the same
>> prefix, then same key in snap_mapper.
>>
>> diff --git a/src/osd/SnapMapper.cc b/src/osd/SnapMapper.cc
>> index 315e2e2..27cc2b7 100644
>> --- a/src/osd/SnapMapper.cc
>> +++ b/src/osd/SnapMapper.cc
>> @@ -72,8 +72,7 @@ string SnapMapper::get_prefix(snapid_
>> t snap)
>>    char buf[100];
>>    int len = snprintf(
>>      buf, sizeof(buf),
>> -    "%.*X_", (int)(sizeof(snap)*2),
>> -    static_cast<unsigned>(snap));
>> +    "%.*llX_", (int)(sizeof(snap)*2), snap);
>>
>> 2015-02-04 12:02 GMT+08:00 Ding Dinghua <dingdinghua85@gmail.com>:
>> > I think so, snap_id is defined and used as uint64_t in ceph, but here
>> > static_cast may introduce bug, since two snap_id may get the same
>> > prefix, then same key in snap_mapper.
>> >
>> > diff --git a/src/osd/SnapMapper.cc b/src/osd/SnapMapper.cc
>> > index 315e2e2..27cc2b7 100644
>> > --- a/src/osd/SnapMapper.cc
>> > +++ b/src/osd/SnapMapper.cc
>> > @@ -72,8 +72,7 @@ string SnapMapper::get_prefix(snapid_
>> > t snap)
>> >    char buf[100];
>> >    int len = snprintf(
>> >      buf, sizeof(buf),
>> > -    "%.*X_", (int)(sizeof(snap)*2),
>> > -    static_cast<unsigned>(snap));
>> > +    "%.*llX_", (int)(sizeof(snap)*2), snap);
>> >
>> > 2015-02-04 1:25 GMT+08:00 Samuel Just <sam.just@inktank.com>:
>> >> Should probably be cast to long unsigned with lX conversion specifier?
>> >> -Sam
>> >>
>> >> On Tue, Feb 3, 2015 at 9:21 AM, Samuel Just <sam.just@inktank.com> wrote:
>> >>> It looks like snapid_t is a uint64_t, but snprintf expects an unsigned there.
>> >>> -Sam
>> >>>
>> >>> On Tue, Feb 3, 2015 at 9:15 AM, Gregory Farnum <greg@gregs42.com> wrote:
>> >>>> On Tue, Feb 3, 2015 at 4:12 AM, Ding Dinghua <dingdinghua85@gmail.com> wrote:
>> >>>>> Hi all:
>> >>>>>       I don't understand why SnapMapper::get_prefix   static_cast snap
>> >>>>> to unsigned:
>> >>>>>
>> >>>>> string SnapMapper::get_prefix(snapid_t snap)
>> >>>>> {
>> >>>>>   char buf[100];
>> >>>>>   int len = snprintf(
>> >>>>>     buf, sizeof(buf),
>> >>>>>     "%.*X_", (int)(sizeof(snap)*2),
>> >>>>>     static_cast<unsigned>(snap));
>> >>>>>   return MAPPING_PREFIX + string(buf, len);
>> >>>>> }
>> >>>>>
>> >>>>> Will this limit snapshot count in pool to 2^32 -1 ?
>> >>>>>
>> >>>>> Could anyone clarify ? Thanks
>> >>>>
>> >>>> I think the code base is a little confused about whether snaps should
>> >>>> be 32 or 64 bits in various places. :( That said, the size of unsigned
>> >>>> can vary across architectures, so this should probably be sized more
>> >>>> explicitly as whatever it's supposed to be on the disk...
>> >
>> >
>> >
>> > --
>> > Ding Dinghua
>>
>>
>>
>> --
>> Ding Dinghua
>> --
>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>>